Using digital twins to prepare a refinery turnaround

A refinery turnaround brings inspection, maintenance, access and production teams into the same planning exercise. A digital twin can help them find the records behind a work item and understand its location. Whether that saves time depends on the quality of the records, the workflow and the people maintaining the model. There is no universal percentage reduction that a project should assume.

This page covers setting up a measurable pilot. For the full planning picture, from scope built on inspection history to access planned in the model and NDT results recorded on the CMLs during the outage, read digital twins for turnarounds.

Start with a decision that currently takes too long

Choose a recurring task: locating a measurement point, retrieving the previous inspection report, checking whether a finding remains open, or assembling the evidence for a planned repair. Record the current process and the time spent on each step. This gives the pilot a defined purpose and a baseline.

Keep the boundary manageable. One unit or asset group with dependable identifiers is more useful for evaluation than a site-wide model whose inspection locations cannot be reconciled. Agree which decisions the model supports and which remain in the existing maintenance or document system.

Connect work items to traceable inspection records

Each work item needs a consistent asset and location reference. Attach the inspection date, method, source report and review status. Keep the original result and its limitations accessible. A planner should be able to distinguish an examined location from an inaccessible area or a location with no usable history.

Reconcile repairs, replacements and changed equipment tags before comparing measurements. If readings were taken at different locations or under different conditions, an apparent trend may be misleading. A visual model can display the evidence, but the responsible technical personnel determine what it supports.

Test the handover between teams

Ask an inspector to locate a finding, a planner to identify the associated work and a reviewer to retrieve the source record. Include an amended report and an unresolved action in the exercise. Check that users see the correct version and understand who can change the status.

Access preparation, isolation, permits and inspection sequencing remain project responsibilities. The model should make relevant constraints easier to find; it does not approve access or authorise continued operation. Record exclusions in the work scope rather than leaving them implicit.

Measure the pilot without inventing a return

Compare the same planning tasks before and after the pilot. Record retrieval time, missing or mismatched records, correction effort and the time required to keep the model current. Separate one-time data cleanup from recurring operating effort. If a different team or a different asset group is used, explain why the comparison is still useful.

Do not convert all saved administrative time into an assumed reduction in shutdown duration. The turnaround schedule may be governed by access, repairs, materials, production constraints or other critical work. A planning improvement has value even when it does not change the completion date.

Define what happens after the turnaround

Assign responsibility for importing final reports, closing findings and updating the asset history. Agree exports, backups, permissions and the link to the system that owns maintenance actions. A model that stops being maintained after the outage will become less useful for the next inspection cycle.

Refinery digital-twin scoping guide · Inspection reporting with asset context · Discuss a pilot using your records